SpaceX (SPCX)

SpaceX combines three quality businesses under one vertical integration: Starlink (the most valuable piece — subscription satellite internet, 63% EBITDA margin, +50% growth — the engine of revenue and future cash), the dominant launch franchise (more than 80% of the world's mass to orbit, whose greater value is internal because it subsidizes Starlink with the lowest-cost access to orbit), and a real AI lab (xAI, Colossus, and ~US$26bn/year of contracted compute). Rocket reusability is the cost advantage that underpins the whole group.

Since the lines are of a different nature, the valuation is a sum of the parts. Each piece is valued on EV/EBITDA — the metric the data supports for a business with negative consolidated FCF: Starlink ~US$459bn (17x EBITDA), Space ~US$83bn (18x), and xAI ~US$300bn (25x), which together add up to ~US$842bn of terminal operating value; net of cash and debt, base operating equity runs around US$829bn over five years — ~US$63 per share (well below the price even brought to today).

Against a price of US$138 (a market cap of ~US$2.05 trillion), the implied return is negative: the base scenario yields a CAGR of -15%. Negative consolidated FCF is shown explicitly — the cash burn is real and depends on financing — not omitted; it's the correct lens for a business in investment phase.

No margin of safety: the price already discounts a demanding scenario. The market price (~US$2.05 trillion) sits well above the sum-of-the-parts operating value across all three scenarios: the base runs around ~US$0.83 trillion, and even the bull scenario (~US$1.65 trillion) falls short of the price. As a result, the return at today's price is negative, and the intrinsic value brought to today at the risk-free rate sits below the price (a premium, not a discount). The verdict is Overvalued: quality businesses, but at a price that discounts flawless execution of the full vision.

The central disconfirming test is financial: the trajectory of consolidated FCF and capex. If capex doesn't moderate and Starlink doesn't accelerate its free-cash-flow generation, dependence on external financing becomes the dominant risk. In parallel, three execution tests: whether Starship reaches payload operations (enabling the next cost step-down and Starlink V3), whether xAI shifts from burning cash to monetizing contracted compute, and whether competition (Kuiper) fails to erode the lead in low-earth-orbit broadband.

Mars, the lunar economy, and orbital data centers at scale remain genuinely non-quantifiable optionalities: they aren't paid for in the base case — they're excess return only if they materialize — unlike Starlink, Space, and xAI, which do carry present value even though the group's consolidated FCF is negative today.